Biomedical Engineering Reference
In-Depth Information
owing to increased secondary structure. Substitution of
nonstandard bases
such
as deazaG for G alleviates this difficulty.
g.
UV-sensitive nucleoside analogues in cell growth for fragmentation
. An-
other approach is to grow immortalized lymphoblast cell lines in the presence of
UV-sensitive nucleoside analogues Pirrung et al. (48). These analogues can be
incorporated into the DNA of the cells, which could subsequently be cleaved by
exposure to UV light. The concentration of the analogues determines the fre-
quency of their incorporation, and the size of the resulting fragments.
3.4. Creation of the Tagged Biomolecular Database
These DNA tags are composed of a concatenation of short subsequences,
which encode scalar data values. For example, the information tags may contain
the individual's unique ID and the cell type (in the case of reverse-transcript
cDNA obtained from the RNA expressed by a particular cell type) of the ge-
nomic DNA and may also encode other useful information (e.g., sex and birth-
date of the individual). The tagging can be done using known methods, for
example, a primer-extension reaction, using the fact that one of the ends of the
genomic cDNA can be predicted by the use of the appropriate initial fragmenta-
tion process, and further designing the tags with ends complementary to these
sticky ends resulting from the fragmentation process. The resultant database
elements have tags on each 5'- and 3'-end. This can be done so that each Bio-
molecular Database strand bears a universal amplification (primer) sequence at
the extreme 5'- and 3'-ends.
3.5. DNA Word Design for the Information Tags
A key problem is the design of a lexicon of short DNA sequences (DNA
words) for the information tags in our Biomolecular Database. (Our DNA "in-
formation tag" sequences are in general a subset of such a lexicon). Careful
word design is crucial for optimizing error control in the queries executed within
the Biomolecular Database. Good word design can be used to minimize un-
wanted secondary structure and to minimize mismatching by maximizing bind-
ing specificity. There are conflicting requirements on word design: as strand
length decreases (which is desirable), the difference between distinct informa-
tion words decreases (not desirable). Prior work in DNA word design includes a
four-base mismatch word design used for surface-based DNA computing (29),
and Frutos et al. (24) showed that surface morphology may be an important fac-
tor for discrimination of mismatched DNA sequences. A three-base design was
used by Cukras and coworkers (17). Evolutionary search methods for word de-
signs are described by Deaton et al. (18). Other DNA word designs are de-
